checkpoint.c

linux 内核源代码

			if (!retry && lock_need_resched(&journal->j_list_lock)){
				spin_unlock(&journal->j_list_lock);
				retry = 1;
				break;
			}
		}

		if (batch_count) {
			if (!retry) {
				spin_unlock(&journal->j_list_lock);
				retry = 1;
			}
			__flush_batch(journal, bhs, &batch_count);
		}

		if (retry) {
			spin_lock(&journal->j_list_lock);
			goto restart;
		}
		/*
		 * Now we have cleaned up the first transaction's checkpoint
		 * list. Let's clean up the second one
		 */
		__wait_cp_io(journal, transaction);
	}
out:
	spin_unlock(&journal->j_list_lock);
	result = cleanup_journal_tail(journal);
	if (result < 0)
		return result;
	return 0;
}

/*
 * Check the list of checkpoint transactions for the journal to see if
 * we have already got rid of any since the last update of the log tail
 * in the journal superblock.  If so, we can instantly roll the
 * superblock forward to remove those transactions from the log.
 *
 * Return <0 on error, 0 on success, 1 if there was nothing to clean up.
 *
 * Called with the journal lock held.
 *
 * This is the only part of the journaling code which really needs to be
 * aware of transaction aborts.  Checkpointing involves writing to the
 * main filesystem area rather than to the journal, so it can proceed
 * even in abort state, but we must not update the journal superblock if
 * we have an abort error outstanding.
 */

int cleanup_journal_tail(journal_t *journal)
{
	transaction_t * transaction;
	tid_t		first_tid;
	unsigned long	blocknr, freed;

	/* OK, work out the oldest transaction remaining in the log, and
	 * the log block it starts at.
	 *
	 * If the log is now empty, we need to work out which is the
	 * next transaction ID we will write, and where it will
	 * start. */

	spin_lock(&journal->j_state_lock);
	spin_lock(&journal->j_list_lock);
	transaction = journal->j_checkpoint_transactions;
	if (transaction) {
		first_tid = transaction->t_tid;
		blocknr = transaction->t_log_start;
	} else if ((transaction = journal->j_committing_transaction) != NULL) {
		first_tid = transaction->t_tid;
		blocknr = transaction->t_log_start;
	} else if ((transaction = journal->j_running_transaction) != NULL) {
		first_tid = transaction->t_tid;
		blocknr = journal->j_head;
	} else {
		first_tid = journal->j_transaction_sequence;
		blocknr = journal->j_head;
	}
	spin_unlock(&journal->j_list_lock);
	J_ASSERT(blocknr != 0);

	/* If the oldest pinned transaction is at the tail of the log
           already then there's not much we can do right now. */
	if (journal->j_tail_sequence == first_tid) {
		spin_unlock(&journal->j_state_lock);
		return 1;
	}

	/* OK, update the superblock to recover the freed space.
	 * Physical blocks come first: have we wrapped beyond the end of
	 * the log?  */
	freed = blocknr - journal->j_tail;
	if (blocknr < journal->j_tail)
		freed = freed + journal->j_last - journal->j_first;

	jbd_debug(1,
		  "Cleaning journal tail from %d to %d (offset %lu), "
		  "freeing %lu\n",
		  journal->j_tail_sequence, first_tid, blocknr, freed);

	journal->j_free += freed;
	journal->j_tail_sequence = first_tid;
	journal->j_tail = blocknr;
	spin_unlock(&journal->j_state_lock);
	if (!(journal->j_flags & JFS_ABORT))
		journal_update_superblock(journal, 1);
	return 0;
}


/* Checkpoint list management */

/*
 * journal_clean_one_cp_list
 *
 * Find all the written-back checkpoint buffers in the given list and release them.
 *
 * Called with the journal locked.
 * Called with j_list_lock held.
 * Returns number of bufers reaped (for debug)
 */

static int journal_clean_one_cp_list(struct journal_head *jh, int *released)
{
	struct journal_head *last_jh;
	struct journal_head *next_jh = jh;
	int ret, freed = 0;

	*released = 0;
	if (!jh)
		return 0;

	last_jh = jh->b_cpprev;
	do {
		jh = next_jh;
		next_jh = jh->b_cpnext;
		/* Use trylock because of the ranking */
		if (jbd_trylock_bh_state(jh2bh(jh))) {
			ret = __try_to_free_cp_buf(jh);
			if (ret) {
				freed++;
				if (ret == 2) {
					*released = 1;
					return freed;
				}
			}
		}
		/*
		 * This function only frees up some memory
		 * if possible so we dont have an obligation
		 * to finish processing. Bail out if preemption
		 * requested:
		 */
		if (need_resched())
			return freed;
	} while (jh != last_jh);

	return freed;
}

/*
 * journal_clean_checkpoint_list
 *
 * Find all the written-back checkpoint buffers in the journal and release them.
 *
 * Called with the journal locked.
 * Called with j_list_lock held.
 * Returns number of buffers reaped (for debug)
 */

int __journal_clean_checkpoint_list(journal_t *journal)
{
	transaction_t *transaction, *last_transaction, *next_transaction;
	int ret = 0;
	int released;

	transaction = journal->j_checkpoint_transactions;
	if (!transaction)
		goto out;

	last_transaction = transaction->t_cpprev;
	next_transaction = transaction;
	do {
		transaction = next_transaction;
		next_transaction = transaction->t_cpnext;
		ret += journal_clean_one_cp_list(transaction->
				t_checkpoint_list, &released);
		/*
		 * This function only frees up some memory if possible so we
		 * dont have an obligation to finish processing. Bail out if
		 * preemption requested:
		 */
		if (need_resched())
			goto out;
		if (released)
			continue;
		/*
		 * It is essential that we are as careful as in the case of
		 * t_checkpoint_list with removing the buffer from the list as
		 * we can possibly see not yet submitted buffers on io_list
		 */
		ret += journal_clean_one_cp_list(transaction->
				t_checkpoint_io_list, &released);
		if (need_resched())
			goto out;
	} while (transaction != last_transaction);
out:
	return ret;
}

/*
 * journal_remove_checkpoint: called after a buffer has been committed
 * to disk (either by being write-back flushed to disk, or being
 * committed to the log).
 *
 * We cannot safely clean a transaction out of the log until all of the
 * buffer updates committed in that transaction have safely been stored
 * elsewhere on disk.  To achieve this, all of the buffers in a
 * transaction need to be maintained on the transaction's checkpoint
 * lists until they have been rewritten, at which point this function is
 * called to remove the buffer from the existing transaction's
 * checkpoint lists.
 *
 * The function returns 1 if it frees the transaction, 0 otherwise.
 *
 * This function is called with the journal locked.
 * This function is called with j_list_lock held.
 * This function is called with jbd_lock_bh_state(jh2bh(jh))
 */

int __journal_remove_checkpoint(struct journal_head *jh)
{
	transaction_t *transaction;
	journal_t *journal;
	int ret = 0;

	JBUFFER_TRACE(jh, "entry");

	if ((transaction = jh->b_cp_transaction) == NULL) {
		JBUFFER_TRACE(jh, "not on transaction");
		goto out;
	}
	journal = transaction->t_journal;

	__buffer_unlink(jh);
	jh->b_cp_transaction = NULL;

	if (transaction->t_checkpoint_list != NULL ||
	    transaction->t_checkpoint_io_list != NULL)
		goto out;
	JBUFFER_TRACE(jh, "transaction has no more buffers");

	/*
	 * There is one special case to worry about: if we have just pulled the
	 * buffer off a running or committing transaction's checkpoing list,
	 * then even if the checkpoint list is empty, the transaction obviously
	 * cannot be dropped!
	 *
	 * The locking here around t_state is a bit sleazy.
	 * See the comment at the end of journal_commit_transaction().
	 */
	if (transaction->t_state != T_FINISHED) {
		JBUFFER_TRACE(jh, "belongs to running/committing transaction");
		goto out;
	}

	/* OK, that was the last buffer for the transaction: we can now
	   safely remove this transaction from the log */

	__journal_drop_transaction(journal, transaction);

	/* Just in case anybody was waiting for more transactions to be
           checkpointed... */
	wake_up(&journal->j_wait_logspace);
	ret = 1;
out:
	JBUFFER_TRACE(jh, "exit");
	return ret;
}

/*
 * journal_insert_checkpoint: put a committed buffer onto a checkpoint
 * list so that we know when it is safe to clean the transaction out of
 * the log.
 *
 * Called with the journal locked.
 * Called with j_list_lock held.
 */
void __journal_insert_checkpoint(struct journal_head *jh,
			       transaction_t *transaction)
{
	JBUFFER_TRACE(jh, "entry");
	J_ASSERT_JH(jh, buffer_dirty(jh2bh(jh)) || buffer_jbddirty(jh2bh(jh)));
	J_ASSERT_JH(jh, jh->b_cp_transaction == NULL);

	jh->b_cp_transaction = transaction;

	if (!transaction->t_checkpoint_list) {
		jh->b_cpnext = jh->b_cpprev = jh;
	} else {
		jh->b_cpnext = transaction->t_checkpoint_list;
		jh->b_cpprev = transaction->t_checkpoint_list->b_cpprev;
		jh->b_cpprev->b_cpnext = jh;
		jh->b_cpnext->b_cpprev = jh;
	}
	transaction->t_checkpoint_list = jh;
}

/*
 * We've finished with this transaction structure: adios...
 *
 * The transaction must have no links except for the checkpoint by this
 * point.
 *
 * Called with the journal locked.
 * Called with j_list_lock held.
 */

void __journal_drop_transaction(journal_t *journal, transaction_t *transaction)
{
	assert_spin_locked(&journal->j_list_lock);
	if (transaction->t_cpnext) {
		transaction->t_cpnext->t_cpprev = transaction->t_cpprev;
		transaction->t_cpprev->t_cpnext = transaction->t_cpnext;
		if (journal->j_checkpoint_transactions == transaction)
			journal->j_checkpoint_transactions =
				transaction->t_cpnext;
		if (journal->j_checkpoint_transactions == transaction)
			journal->j_checkpoint_transactions = NULL;
	}

	J_ASSERT(transaction->t_state == T_FINISHED);
	J_ASSERT(transaction->t_buffers == NULL);
	J_ASSERT(transaction->t_sync_datalist == NULL);
	J_ASSERT(transaction->t_forget == NULL);
	J_ASSERT(transaction->t_iobuf_list == NULL);
	J_ASSERT(transaction->t_shadow_list == NULL);
	J_ASSERT(transaction->t_log_list == NULL);
	J_ASSERT(transaction->t_checkpoint_list == NULL);
	J_ASSERT(transaction->t_checkpoint_io_list == NULL);
	J_ASSERT(transaction->t_updates == 0);
	J_ASSERT(journal->j_committing_transaction != transaction);
	J_ASSERT(journal->j_running_transaction != transaction);

	jbd_debug(1, "Dropping transaction %d, all done\n", transaction->t_tid);
	kfree(transaction);
}